Introduction

Lymphocytes are important mediators of the immune response to cancer [1]. Although the immune system has long been recognized for its vital role in tumor surveillance, the recent advent of immunotherapy has highlighted the importance of preserving a pool of lymphocytes in the circulation [2]. Reductions in total lymphocyte count (TLC) and lymphocyte infiltration in pathologically resected specimens are associated with poor disease-free survival and overall survival (OS; [3,4,5,6,7,8,9,10,11]).

Lymphocytes are the most radiosensitive cells among cells of erythroid, myeloid, and lymphoid lineages [12]. The lethal dose required to reduce the surviving fraction of lymphocytes by 50 and 90% is 2 and 3 Gy, respectively [13], which can affect a substantial proportion of circulating lymphocytes during a course of conventional fractionated radiotherapy (RT; [14]). Therefore, lymphopenia is known to be a common side effect of RT. Radiotherapy-related lymphopenia has been reported in various tumor types such as glioblastoma, pancreatic cancer, and non-small cell lung cancer [3,4,5, 10, 11, 15, 16].

Radiotherapy has been widely performed for patients with hepatocellular carcinoma (HCC) in clinical practice [17, 18] and is recently listed in the National Comprehensive Cancer Network Guidelines [19], but RT-related lymphopenia in patients with HCC has not been well investigated to date [20]. Compared with other types of tumors, HCC is unique with regard to RT-related lymphopenia; it is a hypervascular tumor and its location in the liver harbors a very rich blood circulation. This leads to a greater amount of blood being exposed to radiation, which can maximize the radiation effect and cause lymphopenia. In this study, we investigated the prognostic significance of acute severe lymphopenia (ASL) for HCC patients treated with RT and clinical factors that might predict the development of ASL.

Patients and methods

Patient population

We retrospectively identified patients with HCC who underwent RT between February 2001 and December 2016 at our institution. Eligible patients were required to have TLC, neutrophil–lymphocyte ratio (NLR), platelet–lymphocyte ratio (PLR), and C‑reactive protein (CRP) values at baseline, and at least one TLC value within 3 months of initiating RT. Patients were excluded based on the presence of distant metastasis, other previous or concurrent malignancies, or a history of RT. As a result, 920 patients were included in the final analysis.

Immunological markers

Peripheral blood counts were typically assessed every week during RT, followed by every 1–3 months after RT for 1 year. Total lymphocyte counts were recorded from the clinical complete blood count panel. Total lymphocyte count data were classified based on the months from the start of RT. If TLC was not available at a certain month, the closest value to the month was used. Lymphopenia was graded using the Common Terminology Criteria for Adverse Events version 4.03, where TLC from the lower limit of the normal range to 800/μL was considered as grade 1, 800–500/μL as grade 2, 500–200/μL as grade 3, and <200/μL as grade 4. Acute severe lymphopenia was defined as any occurrence of TLC <500/μL (grade 3/4 lymphopenia) within 3 months of beginning RT, as previously described [5]. The NLR was calculated by dividing the neutrophil count by the lymphocyte count, and the PLR was calculated by dividing the platelet count by the lymphocyte count. The baseline NLR, PLR, and CRP values were measured before the start of treatment within 2 weeks.

Radiotherapy

External beam RT using either three-dimensional (515 patients, 56.0%) or intensity-modulated RT (405 patients, 44.0%) was performed in all patients. The RT type and combined modality with other treatments were determined through multidisciplinary discussion after considering the disease status and the patient’s condition.

Total RT dose was determined by the volume of normal liver and the maximum dose to the stomach or duodenum. When applying conventional fractionated radiotherapy (CFRT), typically 45–60 Gy in 20–25 fractions was used, and a central simultaneous integrated boost of 50–100 Gy in 20–25 fractions was applied with intensity-modulated RT. When applying stereotactic body radiotherapy (SBRT), a total dose of 60 or 52 Gy in 4 fractions was typically used for PTV.

When applying CFRT, combined modality such as transarterial chemoembolization (TACE) or hepatic arterial infusion chemotherapy (HAIC) was often used; patients were considered for additional RT when TACE was considered incomplete, which was defined as incomplete tumor filling by iodized oil based on enhanced computed tomography scans performed within 4–6 weeks of TACE or newly seen enhancing soft tissue in the post-TACE lesion within 3 months of TACE; HAIC was administered in concurrent continuous infusion of 5‑fluorouracil (500 mg/m2/day) during the first and last week of RT using a percutaneous hepatic arterial catheter inserted via hepatic arterial angiography. Tumors with a diameter of ≤5.0 cm for a single tumor or with the sum of the diameters being ≤6.0 cm for up to three tumors were generally indicated for SBRT, assuring 1 cm distance from the organ at risk involving the gastrointestinal tract.

Statistical analysis

The change of TLC over time was analyzed using an analysis of variance (ANOVA) with repeated-measures design. The differences between baseline level and the levels at the respective timepoints were determined using an independent t test. The OS rates were calculated from the start of RT using the Kaplan–Meier method and then compared using the log-rank test. Univariate and multivariate analyses to identify predictors of OS and ASL were performed using Cox regression and logistic regression modeling, respectively. Factors with a p value <0.05 in the univariate analyses were included in the multivariate analysis. All tests were two-sided and significance was defined as p <0.05. Statistical analyses were performed with the Statistical Package for Social Sciences, version 23.0 (IBM SPSS Statistics, Armonk, NY, USA).

Results

Patients

The baseline patient characteristics are summarized in Table 1. The median age was 57 years (interquartile range [IQR], 50–66 years), and the median tumor size was 7.8 cm (IQR, 4.0–11.6 cm). Barcelona Clinic Liver Cancer (BCLC) stage was A in 98 patients (10.7%), B in 220 patients (23.9%), and C in 602 patients (65.4%). The median biologically effective dose calculated with an α/β of 10 was 62.5 Gy (IQR, 53.1–78.0 Gy), and the median planning target volume (PTV) was 606 cm3 (IQR, 229–1335 cm3). In all, 90 patients (9.8%) and 830 patients (90.2%) received SBRT and CFRT, respectively; 536 patients (58.3%) were treatment naïve. There was no significant difference in baseline TLC between treatment-naïve patients and non-treatment-naïve patients (median, 1050/μl vs. 1080/μl; p = 0.826).

Table 1 Baseline characteristics in all patients (n = 920)
Full size table

TLC over time

The change of TLC was significant over time (p < 0.001). When compared with the baseline level, the TLCs at the all monthly timepoints were significantly reduced (all p < 0.001). Before RT, the median TLC for all patients was 1120/μl and the grade 3/4 lymphopenia rate was 5.0%. One month after the start of RT, the median TLC decreased to 310/μl and the grade 3/4 lymphopenia rate was 72.8%. Two months after beginning RT, the TLC partially recovered to a median of 840/μl and the grade 3/4 lymphopenia rate was 23.4%. Three months after beginning RT, the median TLC was 650/μl and the grade 3/4 lymphopenia rate was 31.7%. After that, the median TLC remained similar and did not completely recover to its initial level until the follow-up of 1 year (Fig. 1a and c). In long-term survivors who survived more than 15.8 months, the median OS of all patients, the median TLC was higher at all the measured time points than that of the short-term survivors who survived less than 15.8 months (Fig. 1b). Overall, ASL, defined as any occurrence of TLC <500/μl (grade 3/4 lymphopenia) within 3 months of beginning RT, developed in 804 patients (87.4%).

Fig. 1
figure 1

a Total lymphocyte count in all patients, b total lymphocyte count in patients with overall survival (OS) more and less than 15.5 months, and c the grade of lymphopenia over time

Full size image

Prognostic factors

The median follow-up period was 15.8 months (IQR, 7.3–30.1 months) for all patients and 36.3 months (IQR, 24.8–54.9 months) for patients who were alive at the time of analysis. The median OS for all patients was 15.8 months. The results of the univariate and multivariate Cox regression analyses for OS are provided in Table 2. In the multivariate analysis, ASL (hazard ratio [HR], 1.40; 95% confidence interval [CI], 1.02–1.91; p = 0.035) and the number of tumors, Child–Pugh class, previous treatment history, biologically effective dose, combined HAIC, PTV, and baseline NLR were independent significant prognostic factor of OS (all p < 0.05). The median OS in patients with and without ASL was 13.6 and 46.7 months, respectively, in all patients (p < 0.001; Fig. 2a); 17.9 and 46.7 months, respectively, in the PTV <606 cm3 subgroup (p < 0.001; Fig. 2b); 11.7 and 48.7 months, respectively, in the PTV ≥606 cm3 subgroup (p = 0.041; Fig. 2c); 22.5 and 34.4 months, respectively, in the BCLC-A subgroup (p = 0.183; Fig. 2d); 20.6 and 46.3 months, respectively, in the BCLC-B subgroup (p = 0.009; Fig. 2e); and 11.0 and 46.7 months, respectively, in the BCLC-C subgroup (p < 0.001; Fig. 2f).

Table 2 Analysis of factors affecting overall survival
Full size table
Fig. 2
figure 2

Overall survival of patients with and without acute severe lymphopenia (ASL): a all patients, b planning target volume (PTV) <606 cm3 group, c PTV ≥606 cm3 group, d BCLC-A group, e BCLC-B group, and f BCLC-C group. BCLC Barcelona Clinic Liver Cancer stage

Full size image

Predictors of ASL

The results of univariate and multivariate logistic regression analyses for ASL are provided in Table 3. In the multivariate analysis, larger PTV (per 10 cm3: HR, 1.02; 95% CI, 1.01–1.03; p < 0.001), lower baseline TLC (per 100/μl: HR, 0.86; 95% CI, 0.82–0.91; p < 0.001), and use of HAIC (HR, 2.26; 95% CI, 1.08–4.75; p = 0.031) were independently associated with an increased risk of ASL, while hypofractionation (SBRT: HR, 0.19; 95% CI, 0.07–0.49; p = 0.001) was independently associated with a decreased risk of ASL. When patients were stratified according to these risk factors, patients with high baseline TLC and SBRT with small PTV had the lowest incidence of ASL (16 of 61 patients; 26.2%), while patients with low baseline TLC, CFRT with large PTV, and combined HAIC showed the highest incidence of ASL (261 of 264 patients; 98.9%; Fig. 3).

Table 3 Analysis of factors affecting acute severe lymphopenia
Full size table
Fig. 3
figure 3

Acute severe lymphopenia (ASL) rates and median TLC according to the risk groups. CFRT conventional fractionated radiotherapy, HAIC hepatic arterial infusion chemotherapy, SBRT stereotactic body radiotherapy, TACE transarterial chemoembolization, TLC total lymphocyte count

Full size image

Discussion

This study showed that ASL was an independent risk factor that predicted poor survival. Radiotherapy-related parameters, such as treatment volume and number of fractions, as well as combined modality and baseline TLC were independently associated with the development of ASL. This is one of the first studies that identified the occurrence of lymphopenia after RT for HCC [21]. In addition, to our knowledge, this is the largest retrospective study among various types of solid tumors that examined RT-related lymphopenia.

Although radiation is known as local modality, RT to peripheral organs can systemically affect a large proportion of blood volume during a course of multiple fractionation. Lymphocytes are the most radiosensitive cells among all blood cell types and some lymphocyte subpopulation can be more sensitive to radiation. Recent ex vivo/in vitro studies showed that a single dose of 2 Gy or more can indue apoptosis and necrosis of lymphocytes, and even the lower doses of 0.3–0.7 Gy can affect the more radiosensitive natural killer (NK) and B cells [22]. In addition, other recent in vitro studies showed that ATM(Ataxia-Telangiectasia Mutated)-expressed lymphocytes are highly radiosensitive [23]. Accordingly, the effect of radiation on the circulating blood volume can easily induce depletion of lymphocytes. In a prospective clinical study, a limited abdominal target volume with 14–26 Gy doses in a standard fractionation induced a profound and prolonged depletion of lymphocytes [24]. However, to analyze how the human immune system copes with a loss of lymphocytes comprehensively, the functionality of the remaining lymphocytes should be taken into account. The detailed immunophenotyping of blood can be a viable approach to evaluate the functional state of the lymphocytes in a future examination [25].

In these patients, lymphopenia peaked at 1 month and partially recovered after 2 months. However, during a full year of observation, TLC remained persistently low and did not recover to its baseline level. Grossman et al. also reported that TLC in patients treated with RT for solid tumors, such as malignant glioma, pancreatic cancer, and lung cancer, remained low and did not recover after 1 year [3]. Likewise, the CD4 count after applying RT and temozolomide to patients with high-grade gliomas showed a similar trend [4]. In addition, chronic lymphopenia after RT can last for years according to previous reports [15, 26]. Our study also confirmed that the pattern of chronic lymphopenia in patients with HCC is consistent with that in patients with other tumor types.

In this study, patients with ASL had significantly shorter OS (median, 13.6 vs. 46.7 months), and ASL was an independent factor associated with poor OS after adjusting for other factors. A possible explanation is that the immune suppression of these patients could compromise survival. It has long been thought that the tumor-infiltrating lymphocytes play a significant role in controlling the growth of cancers [8]. Histological evidence of T cells infiltrating tumor islets has been associated with improved survival [27,28,29]. In addition, reduced lymphocyte counts and reduced lymphocyte infiltration in pathologically resected specimens have been proven to have a negative impact on OS among various tumor types in a clinical setting [3,4,5,6,7,8,9,10]. Given that circulating lymphocytes are the cells that eventually infiltrate tumors, RT-induced lymphopenia in peripheral blood might contribute to suboptimal treatment outcomes through the lack of immune reaction. Another possibility is that ASL is a surrogate representing the patient’s overall health status, rather than a direct cause for decreased OS. Interestingly, the significant predictors of ASL in the univariate analyses were more or less the same as the significant prognostic factors of OS in the univariate analyses (Tables 2 and 3).

Because ASL has a great impact on prognosis, we attempted to elucidate the factors causing it. In this study, larger PTV, lower baseline TLC, and the use of HAIC were independently associated with an increased risk of ASL, while hypofractionation was independently associated with a decreased risk of ASL. The association between RT-related parameters and ASL can be supported by a mathematical computation model [14]. The model determined that in a typical RT plan for glioblastoma, a single radiation fraction delivered 0.5 Gy to 5% of circulating cells; after 30 fractions 99% of circulating blood had received ≥0.5 Gy. Moreover, the model examined two different PTVs (4.2 cm3 and 268 cm3) and showed a large difference in the proportion of irradiated blood. Previous studies with smaller sample sizes also reported that a large PTV was associated with increased incidence of ASL in 210 patients with glioblastoma [16], and SBRT (5 fractions) was less associated with a decrease in TLC compared with CFRT (28 fractions) in 133 patients with pancreatic cancer [10]. Compared with these reports, our study used a larger dataset of more than 1000 patients and proved the theoretical hypothesis on the relationship between RT parameters and TLC clinically. In addition to the RT-related parameters, HAIC was shown as an independent risk factor of ASL. Although HAIC aims to deliver a chemotherapeutic agent to liver only, its systemic myelotoxic effect has been reported [30]. It may also affect lymphocytes in rich blood volume of the hepatic circulation. Therefore, more concern for lymphopenia seems to be warranted when RT is combined with HAIC.

Our results showed that ASL developed in 87.4% of the patients, which was higher than previously reported in other studies [4, 5, 10, 16]. This difference can be explained by the larger tumor sizes and PTV in this study; the median tumor size was 7.8 cm with a median PTV of 606 cm3. In a previous study of RT for glioblastoma, ASL rates were 15.5 and 33.8% with a median PTV of 246 cm3 and 375 cm3, respectively [16]. In another study of RT for pancreatic cancer, ≥grade 3 lymphopenia at 2 months occurred in 13.8 and 71.7% of the patients, with a median PTV of 89 cm3 and 345 cm3, respectively [10]. In addition to the large PTV in this cohort, since HCC is hypervascular and harbors large amounts of blood, the blood exposure to radiation can be maximized. Therefore, it would be necessary to investigate how ASL occurs in this circumstance of maximized exposure of blood to radiation.

To reduce the fractionation number so as to avoid ASL, hypofractionation or SBRT are valuable approaches. To apply hypofractionation or SBRT in various clinical situations, the isotoxic SBRT regimen, developed at the Princess Margaret Hospital at the University of Toronto, can be a useful approach [31]. According to this concept, the dose per fraction was determined based on the effective volume of normal liver irradiated and can range from 5 to 9 Gy in 6 fractions. In the current study, patients with low baseline TLC were at a high risk of developing ASL (Fig. 3). In this particular patient group, more effort is required to reduce the number of fractions to avoid ASL.

The limitations of this study were its retrospective nature and its long study duration, showing heterogeneity in the selection of RT type and combined modalities. In addition, data on incidental irradiation of lymphoid organs, such as the spleen and bone marrow, were not collected. Therefore, our findings need to be interpreted with some caution.

Conclusion

The safety and efficacy of RT for HCC has been shown in many previous studies [32]. In this context, our findings may help establish optimal strategies for RT for HCC. Although lymphopenia after RT is a well-known phenomenon, its clinical implication has not been well defined. However, the antitumor role of lymphocytes has been more emphasized in the modern immunotherapy era. Recently, preliminary data from the CheckMate-040 trial suggested that nivolumab has clinical activity and is tolerable in patients with HCC [33]. Considering this potential synergistic effect between RT and immunotherapy, maintaining and restoring TLC can be key in improving oncologic outcomes. In addition, RT has an immunostimulatory effect via tumor necrosis, the release of tumor-associated antigens, and recruitment of immune cell into the tumor microenvironment [34]. However, RT also has an immunosuppressive effect, which includes the depletion of the lymphocyte pool. Thus, to efficiently exploit the therapeutic potential of RT, optimal strategies need to be identified to preserve the lymphocyte pool after RT. For example, as shown in our findings, alternative radiation regimens, particularly hypofractionated regimens or SBRT, could be more actively administered for the purpose of immune sparing. Furthermore, optimal beam arrangements and field size, which minimize radiation to blood reservoir or lymphoid organs, can be considered. There is growing evidence that RT may affect lymphocyte homeostasis via cytokines, such as interleukin-7, which plays a role in the development of B and T lymphocytes [35,36,37]. Approaches such as the use of exogenous interleukin-7 or reinfusion of previously harvested circulating lymphocytes after therapy could be investigated [38]. These efforts will provide clues to improve survival in patients with HCC, and practicing clinicians and researchers should begin to focus on finding optimal strategies to reduce RT-related lymphopenia.